Smith et al worry that we have underestimated the role of genetics in early-onset depression. They draw our attention to the published evidence for the importance of genetic factors in prepubertal depression, which itself is a marker for adult bipolar disorder, although not necessarily a marker for major depression (the topic of our editorial). From their own data, they report that three-quarters of 47 subjects who had a depressive episode by a mean age of 15 had a first-degree relative with an affective disorder. Strong evidence indeed of familiarity, but not necessarily of a genetic cause. Despite their certainty, many of us have problems with the precise nature of the evidence supporting genetic factors in major depression, in part because of the dimensional nature of depression, and in part because of the extensive comorbidity.

We opined that the heightened risk of depression in young people whose parents had depression was likely to be ‘more psychological than genetic’ and referred the reader to the review by Beardslee et al (Reference Cervilla, Prince and Joels1998). We provided evidence that intervention programmes for adolescents can reduce by half the probability of depression in the future. Smith et al argue that universal interventions are unlikely to be effective until we have complete understanding of how genetic and non-genetic factors interact to bring about the depressive phenotype, and that interventions targeted to high-risk groups should be deferred until we can reliably identify those individuals at high risk.

We strongly oppose this thinking. Most interventions in medicine are introduced before there is a complete understanding of the aetiology of the disorder, and usually before there is precise information as to which individuals will respond. Simply to know that an intervention can produce a reliable and significant benefit is sufficient to warrant implementation. We believe that this is the situation in regard to the prevention of major depression in young people.